You can do all those things meaningfully: I think it's really about the end product (physical or not) - if you are making something you don't believe is useful to society, or don't have control over how your efforts are being used downstream, then it can easily become inherently meaningless.

There are basic needs: clean water, clean air, infrastructure, transportation, logistics, etc., etc. And also "non-tangible ones". You can get much more authentic social respect if you work on products/services that people unambiguously like and need.

I recommend, for example: - Biomedical: signals, images, ... (I've done a bit) - Anything GIS, urban/transportation planning, geo-spatial analytics, cities, etc. (my chosen specialty, very fulfilling).

There's lots of number crunching, but also human-entry string processing (fuzzy matching, etc.). Actually, very versatile programmatically. And then there is fast graphics (OpenGL, etc.) - not my favourite part, actually, but you can outsource it partly.

You get to work in very multi-disciplinary groups, so you can really assess where you want to go long-term. I was surrounded by people with very similar training to mine - technologically it was pretty good, but topic-wise it was a bit of an echo-chamber.